The number of lakes supporting accumulations of toxic cyanobacteria in the United States is rising due to changes in land use and climate. As a result there is increasing risk associated with lake recreational activities (e.g. swimming, boating), and consumption of fish. Physicochemical factors leading to cyanotoxin production are ill-defined at time-scales relevant to cyanobacterial ecology (i.e. minutes to hours). Acute poisonings are well-documented, but chronic exposure to low levels of cyanotoxins in drinking water is not. Lakes in the U.S. provide source water for drinking water treatment plants (DWTP) serving hundreds of millions of people. Lake conditions, or immunological variables, that support cyanotoxin production are intrinsically linked to the occurrence of cyanotoxins in drinking water because these variables also influence cyanotoxin removal efficiency by DWTP processes. The investigators propose to use high-resolution sensors on buoys and a new automated sampling device to investigate relationships between immunological variables and the presence of cyanotoxins in lakes and drinking water. For this work, Lake Winnebago in the Lake Michigan watershed and a DWTP drawing water from the lake will serve as a model system. In aim 1 the investigators will deploy an instrumented buoy and automated sampling device near the DWTP intake. Buoy sensors will measure physical variables and algal pigments while nutrients, community composition, and cyanotoxins will be measured in preserved water samples collected by the automated sampler. Cyanotoxins will be measured by liquid chromatography tandem mass spectrometry and include hepatotoxins (microcystins and cylindrospermopsin) and neurotoxins (anatoxins, saxitoxin, and beta-N-methylamino-L-alanine). In addition, a zebrafish assay will be used to detect unknown toxins and overall water toxicity. Thus, cyanotoxins and other limnological variables will be measured at high resolution (minutes to hours) for the duration of the cyanobacterial growth season (June-October). This data will be used to test fundamental hypotheses about the physiological ecology of cyanotoxin production in lakes. In aim 2, the investigators will deploy automated samplers at a DWTP, collecting samples twice daily. They will explore temporal dynamics of cyanotoxin occurrence in these samples as well as removal efficiency by DWTP processes. In aim 3 the investigators will construct models of cyanotoxin occurrence in lakes and drinking water. The overall goal of this final aim is to produce predictive models of cyanotoxin levels that can be used by lake managers and DWTP operators to prevent human exposure to cyanotoxins. This proposal leverages the support of other ongoing projects including the NSF Undergraduate Research in Biology and Mathematics (UBM) program at the University of Wisconsin, Milwaukee (UWM), and resources, facilities, and services provided by the Global Lake Ecological Observatory Network and the Children's Environmental Health Sciences Core Center at UWM.